Download PDF

An economic evaluation of the osmoregulation gene
technology to the Australian wheat industry
R.J. Farquharson, J.M. Morgan and J.P. Brennan
NSW Agriculture, Tamworth Agricultural Institute, RMB 944 Tamworth NSW 2340,
& Wagga Wagga Agricultural Institute, Private Mail Bag, Wagga Wagga NSW 2650
Abstract
Episodes of rainfall irregularity and soil moisture deficit have focused attention on the
widespread limitation of water supply on winter cereal crop production in Australia.
This has motivated a number of efforts at breeding for improved drought tolerance. A
recent example involves a cellular adaptation which mitigates water loss through
solute accumulation (osmoregulation or osmotic adjustment). An assessment of the
performance of osmoregulation yield response in the presence of climate change
found that wheat cultivars with this gene are unlikely to be adversely affected by
hotter and drier conditions across the wheat belt. The results of an economic
evaluation of potential future innovations (wheat cultivars) from the osmoregulation
gene technology are that for Australia the net present value could range from $388
million to $3.6 billion, depending on the adoption of wheat cultivars with the gene.
Associated benefit-cost ratios ranged from 43:1 to 390:1, and internal rates of return
were 16% to 27%. Even under pessimistic assumptions the returns are quite healthy.
For NSW-only adoption, internal rates of return ranged from 10% to 22%. This
osmoregulation technology has the potential for inclusion in wheat cultivars bred for
other purposes, and for other crops. As such it has implications for agricultural plant
breeders and farmers both in Australia and overseas, and there are substantial
potential spillover benefits.
Key words: Economic evaluation, wheat breeding, osmoregulation
Contributed paper to the 48th Annual Conference of the AARES
11-13 February 2004, Melbourne
Introduction
Episodes of below average rainfall, as in 2002, focus attention on the widespread
limitation of water supply on winter cereal production in Australia. This has
motivated a number of efforts at breeding for improved drought tolerance. One recent
example involves a cellular adaptation which mitigates water loss through solute
accumulation (osmoregulation or osmotic adjustment). Osmoregulation is a response
to water stress, involving intra-cellular solute accumulation in response to a fall in
extra-cellular water potential (Morgan 2002).
Work on osmoregulation began at Tamworth, NSW, with the discovery of large
genotypic differences in wheat for this trait (Morgan 1977). Subsequent work
examined a possible physiological basis for yield effects (Morgan 1980) and
measured yield effects in the field (Morgan 1983). These were found to be dependent
on the interaction of evaporative demand and soil water supply (Morgan 2000).
Evidence of simple inheritance, present in the initial findings, was confirmed in more
formal analyses (Morgan 1983, 1991). A single gene and chromosomal location was
identified (Morgan 1991) and the arm position estimated (Morgan and Tan 1996).
Application to commercial plant breeding was evaluated, with development of a
simple pollen test for the osmoregulation (or) gene (Morgan 1999) and release of
Mulgara, the first commercial cultivar bred specifically for the or gene. Investigations
of breeding application involving evaluation of the technology in a large number of
field experiments – 187 over a period of 9 years in addition to a series of rainshelter
experiments at the one location over 4 years in northern NSW (Morgan et al. 1986,
Morgan 1995, 2000).
Advantages of osmoregulation are apparent when water demand exceeds water
supply, especially in dry periods. Given recent information relating to projected
climate change, assessments of the potential contribution of osmoregulation to
agricultural production have been evaluated for a future climate scenario in the wheat
growing period of the year. These have been examined using a simple model (Morgan
2000), and the implications for this analysis were used to derive likely yield impacts
in the presence of climate change.
This paper contains an evaluation of the potential benefits from using the or
technology in wheat, where the gene is known to be present in a number of cultivars.
Genotypic differences in osmoregulation have also been found in other crop species,
such as sorghum, chickpeas, soybean and barley (Morgan 2003).
A key issue is the potential for the single or gene to be used in breeding new wheat
cultivars, and confer drought tolerance along with other characteristics for which they
might be selected. There is considerable potential from use of this technology,
depending on the frequency of current cultivars in which the gene is not present.
Around 77% of current cultivars are in this category; these are the focus of this paper.
The paper proceeds by investigating potential yield increases within NSW and then
extending the analysis to the whole of Australia. Annual potential benefits based on
full adoption are estimated for each case and then the analysis considers the effects of
different adoption patterns. The resulting benefits are set against the costs of
technology development, and financial measures are calculated.
2
Economic theory context
In terms of neo-classical economic theory and the welfare gains from new
technologies (Alston et al. 1995), the approach used here assumes that supply is
perfectly inelastic and demand perfectly elastic. Extra yield and production due to the
technology are valued at the same price and all benefits are assumed to accrue to
producers. This approach excludes the welfare triangle, but is considered to be a
reasonable representation of the industry welfare change for a technology such as
osmoregulation. Alston et al. (1995) concluded that this method generally provides a
reasonable approximation of total industry benefits from innovations.
Methods of estimating yield changes
The method of evaluation in this paper involves: (a) estimating yield advantages
associated with or for climatic averages of wheat growing regions throughout NSW
and Australia (Commonwealth of Australia, Bureau of Meteorology 2003); (b)
making judgements about the likely extent and rate of adoption of the technology, and
the time period for which it is expected to last; (c) quantifying aggregate benefits
based on a set of prices; and (d) setting these benefits against the relevant costs, so
that appropriate financial return measures can be derived.
A modelling analysis for Australia
The analysis utilises regions defined and used by the Grains Research and
Development Corporation (GRDC) for description and categorisation of wheat
production (ABARE 1999). The yield increases for the various regions were
determined using the model and results in Morgan (2000), as well as from additional
sites. The model requires soil water holding capacities, mean temperatures,
evaporation, rainfall to calculate phenology, and the ratio of evaporative demand to
soil water supply for the period of net positive growth. This ratio is then used to
estimate the relative yield response (yield of lines with high osmoregulation divided
by yield of lines with low osmoregulation) using an empirical relationship based on
field experiments in 5 seasons. Within each region the sites used to estimate response
were limited by availability of appropriate climatic averages. The analysis assumes a
uniform optimal sowing time of May 13 based on Gomez-Macpherson and Richards
(1995) and maximum available soil water at sowing. Because the benefits from this
work will be felt into the future, and in light of predicted future changes in climate, an
analysis is first conducted to look at the implications of these possible changes.
The climate change scenario in CSIRO (2001) predicts hotter and drier conditions in
the future across the wheatbelt, and the issue is how such changes might affect crop
water stress and yield for a genetic technology based on improved osmoregulation
within the wheat plant. The climate inputs for the osmoregulation yield model
(Morgan 2000) of rainfall, evaporation and mean daily temperature were provided for
the reference year (1990) by the Bureau of Meteorology for 15 sites across Australia.
The increases in temperature (mean 1.05oC) and decreases in rainfall (mean 5.4%) are
approximate mid-range values obtained from CSIRO projections for winter and spring
2030 (CSIRO 2001). Evaporation increases (mean 3.8%) were derived from the
projected annual average change in the moisture balance.
Although the projected climate change produces hotter and drier conditions across the
wheatbelt, its effect on crop water stress and osmoregulation yield response was
negligible compared to the geographic variation (Morgan 2003). In fact, of the
3
responsive sites slight reductions occurred with climate change at all except Merredin,
Wongan Hills and Ongerup, where responses increased. Greater reductions in winter
rainfall are predicted in the WA wheatbelt (and a sustained reduction has indeed
occurred since 1970). Overall the mean estimated yield increase for 2030 (19%) was
virtually the same as for 1990 (20%).
The surprising absence of a climate change effect occurs because, with winter crops,
increased temperature hastens development. In general, a 1 degree increase in
temperature reduces time from sowing to flowering by 9 days. This reduces the incrop evaporation substantially more than the rainfall and it approximately offsets the
projected increase in evaporation and reduction in rainfall caused by climate change.
Earlier flowering times would be agronomically possible because the period of frost
risk would also be reduced, possibly by a similar time period.
In view of the absence of an effect of climate change, the economic analysis may be
performed on either actual experimental data or modified responses for varying
locations in the wheatbelt.
Experimental and modelled responses for NSW
The experimental dataset was for the northern wheat belt of NSW. It comprised two
groups of 56 and 131 experiments for the periods 1980–1983 and 1997–2001,
respectively (see Table 1). The results for the first set were published in Morgan et al.
(1986) and the second in Powell (2002). The second publication deals with NSW
Agriculture annual regional trials, which include the cultivars Mulgara and Sunco.
Mulgara (line JM73) is a triple backcross derivative of Sunco, having been bred by
selection for high osmoregulation using pollen grain expression (Morgan 1999). The
yield comparison provides a reasonable estimate of the or gene effect (Morgan 2000).
The modelled data from Morgan (2000) for long-term average conditions used 7
locations in the NSW wheatbelt. This gave a long-term average yield advantage of
12%, broadly in agreement with but somewhat greater than the experimental results of
8% at 3.5 t/ha for the northern wheatbelt (Table 1). However, this value of 12% is
appropriate for the experimental results based on a lower average yield of 2.07 t/ha
for 2000-01 and 2001-02 in Table 2 (Morgan 2004). The average wheat yield in NSW
was 1.94 t/ha in the 1990s decade (source: AWB website:
www.awb.com.au/AWBL/Launch/Site/AboutAWB/Content/CommunityEducation).
Without breeding for the gene, it occurs in only approximately 23% of Australian
varieties (Morgan 2001). There is therefore considerable scope for improvement in
this gene.
Without-project scenario
An important issue in evaluating returns from R&D is the recognition of, and
accounting for, what is likely to have happened if the work had not been done, i.e.
whether the same advances would have been made by other workers perhaps at a later
date. This is the definition of a ‘without-project’ scenario. Marshall and Brennan
(2001) discussed evaluation of the economic impact of a single research project in
terms of its costs and benefits in what is normally a broader research effort. In
particular they were concerned about the inputs from other projects that contribute to
the achievement of advances in the particular research project. These can be termed
technological or farmer-generated ‘spill-ins’.
4
In the case of the or gene, the R&D has comprised a broad program of work over
many years, as outlined above. It is conceivable that someone else could have
conducted such a program at some other time, but there is no evidence of a long term
program involving breeding for this trait in Australia or elsewhere in the world. The
key prerequisites for commercial breeding, namely genetic differences, yield
relationships (including physiology), genetics and linkages, and a simple method of
gene identification are all original work conducted at Tamworth. Therefore the
without-project scenario is that the work would not have been done and the total
potential industry benefits estimated here accrue to the or R&D program.
Benefit estimation and adoption patterns
The approach to benefit estimation involved using model-predicted field-level yield
increases and farm-gate prices to estimate improved returns per ha for different
regions of NSW and Australia. There are no extra farm-level costs associated with
growing the wheat cultivars which include the or gene. These per ha financial benefits
were then aggregated to represent NSW and national financial benefits at full
adoption. Further assumptions about possible adoption patterns were also imposed.
The expected production increase within NSW, based on the average yield for the
period 2000-02, the expected full yield increase (12%) for average climate conditions
(with no climate change impact), and adjusted for the gene frequency (assuming no
intentional breeding for the gene) is estimated to be 665, 000 t, as shown in Table 2.
This is based on complete adoption of the or technology in NSW.
At the national level, the GRDC agro-ecological zone classification for Australian
crop production (ABARE 1999) was used as a detailed wheat industry representation
for benefit estimation, since the yield impacts of the or gene have been shown to vary
with climate and region. Table 3 contains estimates of wheat area harvested, wheat
production and average wheat yield for the average of the years 2000-2002. These
figures were used as a basis for benefit estimation in future years of wheat varieties
containing the or gene.
The model developed from NSW data was used to predict average wheat production
impacts due to the or gene for locations throughout the Australian wheatbelt. At
various sites in the Australian wheatbelt information on soil type, water holding
capacity and the ratio of evaporation to soil water supply was used to calculate the
percentage yield increase due to the osmoregulation technology. These percentage
changes were adapted to the GRDC Agroecological Zones and are also shown in
Table 3. These percentage changes are larger that the 12% for NSW because the
biggest responses to or are in drier areas of SA, Victoria and WA.
When applied to the production figures in the table and after adjusting for or gene
frequency in other cultivars, the potential annual increase in aggregate production at
full adoption is in the order of 2.5 million tonnes of wheat across the whole country.
Extra wheat production in the field was valued at farm-gate prices to generate on-farm
financial benefits. Estimates were made of the average proportion of wheat production
by pay grade in Australian states over the 5-year period to 2001, and these are
presented in Table 4. Wheat prices for December 2003 ($/t FOB, GST exclusive)
5
from the AWB website (http://www.awb.com.au) for Prime Hard ($240), Hard
($229), Premium White ($221), ASW ($209) and Other ($220) were used, with an
average $60 deduction for transport and FOB costs. Information on the wheat grades
grown in each state is shown in Table 4. The price used to value extra production in
each state and region was calculated as a weighted average of the above prices and the
wheat grade proportions in Table 4.
R&D costs
This program of work has not been conducted by a large R&D group with associated
infrastructure and other costs. The development and testing of the or gene and the
associated release of a new variety and pollen test has been largely conducted since
1975 by one officer with technical support and NSW Agriculture inputs to trials in
various locations. Industry funds have been only a minor contribution.
Past costs were derived using estimated full-time equivalent personnel numbers
valued at current salary and on-cost rates. Other inputs were developed based on
recollections of the scale and source of inputs used. The resulting amounts were
adjusted by a GDP deflator and time preference of 4%. Similarly, future inputs to
2030 were estimated for trialling and extension programs; these were also discounted.
Future costs of cultivar development were equivalent to one technical officer plus oncosts and operating ($10 000 per annum) per mainland state from 2004 to 2030. These
costs should not be high because of the existence of the pollen test which can be
carried out at the same time as cross fertilisation. Also, with a single gene, as material
is produced with higher gene frequency there is less need to test for parentage. The
net present value of all these inputs was estimated to be $9.3 million and $4.7 million
for Australia and NSW in year 2003 terms.
Adoption patterns
The question for R&D benefit estimation also includes the issue of representing
possible future adoption patterns, assuming that the predicted yield benefits will lead
to demand for new cultivars and uptake of the technological innovation. The adoption
of wheat cultivars which include the or gene is likely to vary according to climate;
their value will be greater in drier areas.
The analysis used 2003 as the base year for financial calculations. All potential
benefits were assumed to terminate in 2030. Beyond this time the basis for wheat
production in Australia was considered to be sufficiently uncertain that further
quantification was not warranted.
There is evidence of the lag between agricultural research project initiation and
adoption of innovations being a minimum 10-year period (Cox et al. 1997, Marshall
and Brennan 2001). Due to the development work already carried out the or
technology is now available to wheat breeders, but the time period for breeding,
trialling and releasing a new cultivar is likely to be of the order of 5-6 years (the
Mulgara cultivar was released after 8 years). In light of this, two lag periods to
adoption were used – 5 and 10 years.
Marshall and Brennan (2001) discussed how R&D projects contribute to advances
and innovations. The project contributes to an advance which can provide an input to
6
the production of innovations. Their point was to ensure that the costs of developing
the innovation are included when evaluating the potential benefits from an R&D
project. In the present case the advance is the isolation and confirmation of the or
gene (including the development of a pollen test), and the innovation is the
development of cultivars (like Mulgara) adapted to particular regions in Australia. We
have tried to incorporate both the past costs of achieving the advance and the likely
future costs of developing innovative wheat cultivars throughout NSW and Australia.
A linear adoption pattern was utilised (Marshall and Brennan 2001), which was
characterised by a lag to commencement of adoption, a period of linear adoption and
a maximum (or ceiling) level of adoption. There was no limit placed on the period for
which the technology was maintained at its ceiling level beyond the overall time
restriction of 2030. The nature of the or technology is that it is a single gene which,
once selected, can be maintained indefinitely while other genetic advances can be
pursued. Hence the or technology will not be superseded in the current breeding and
genetic framework for developing wheat cultivars.
Scenarios evaluated were lags to adoption of 5 and 10 years, time to adopt of 5, 10
and 15 years, and levels of adoption of 20%, 50% and 80%.
Economic results
If the or gene can potentially be incorporated into wheat varieties that are adaptable in
NSW and throughout Australia then the possible benefits from this technology are
very large. The results of the economic evaluation reflect this, especially since the
R&D costs are quite small. There are no extra farm production costs assumed for
cultivars which might be developed with this gene.
The main results are shown in Table 5, where the Net Present Value is presented (in
units of 2003 $’million) for both NSW and Australia at a real discount rate of 4%.
Depending on assumptions for lag to adoption, time to adopt and level of adoption the
NPV ranged from $101 million to $958 million for NSW, with associated BCRs of
22:1 and 204:1 and IRRs of 10 and 22%, respectively. For Australia the NPVs ranged
from $388 million to $3.6 billion, with associated BCRs of 43:1 and 390:1 and IRRs
of 16% and 27%, respectively.
In these results the NPV and BCR figures appear to be relatively high compared to the
IRRs. The reason for this is that the period of R&D inputs is relatively long compared
to the later (shorter) period of annual benefits. This is a characteristic of the IRR
measure and is a reason why all three financial return measures should be considered.
Two sensitivity analyses of these results were performed for Australia. They involved
using a different base for the production figures used to project future benefits, and
considering the effects of reducing the potential yield advantage from the or gene
from 12% to 8%. The 8% yield increase was from the measured changes from
experimental data presented in Table 1. The production figures in Tables 2 and 3,
which were averages for the first years of the 2000 decade, were replaced with
averages from the 6 years 1992-93 to 1997-98, which was the original data series
presented in ABARE (1999) for the GRDC agroecological zones. For all Australia the
average wheat production was reduced from 22.027 million tonnes (Table 3) to
15.704 million tonnes, and the average Australian yield was reduced from 2.1 t/ha to
7
1.8 t/ha. These lower figures provide a more conservative estimate of the potential
benefits from the technology.
The sensitivity results presented in Table 6 indicate that for Australia when 1990s
production figures are used as a base, the NPV varied from $263 million to $2.5
billion, with associated BCRs of 29:1 and 268:1 and IRRs of 14% to 25%,
respectively. With a lower yield increase and the 2000s production figures, the NPVs
ranged from $206 million to $2.0 billion, with associated BCRs of 23:1 to 212:1 and
IRRs of 14% to 24%, respectively.
Discussion
Potential benefits from the advance in plant breeding associated with the or gene
technology are substantial if innovative wheat cultivars can be developed in the
future. The NPV figures from this evaluation are substantial if the technology is
adopted at high levels across Australia, where a 12% yield improvement based on
current production and yield levels could provide industry benefits from $388 million
to $3.6 billion in 2003 dollar terms. The associated BCRs are very large, since the
costs have been relatively small, and the IRR figures range from 16 to 27%. At lower
levels of potential yield improvement and lower production levels as a base for
projection the Australian results are still very positive.
Even if the benefits are confined to NSW the figures are still very healthy. The most
pessimistic adoption and yield improvement scenario in NSW still returns an NPV of
$101 million, a BCR of 22:1 and an IRR of 10%. These are substantial figures for
essentially the work of one professional officer.
These results can be considered in light of Alston et al. (2000) who considered ranges
of rates of return from a large sample of evaluation studies. For research-only studies,
their results showed that (excluding outliers) the distribution of IRRs had a mean of
79.6%, a mode of 26% and a median of 49%. The above results for Australia compare
favourably with this distribution. The IRR figures in this analysis may be affected by
long costing periods.
Another issue for this technology is that there are substantial spillovers of benefits to
wheat breeding in other countries and to other crop species (sorghum, chickpeas,
soybeans, barley and Brassicas). These effects have not been included in this analysis
and, as Alston (2002) has noted, this is likely to have underestimated the returns from
this R&D program substantially.
We conclude that a relatively conservative evaluation of the or gene technology, if
used to generate new wheat cultivars across a variety of geographical regions in
Australia, is likely to earn substantial industry benefits from the investment in the
particular R&D program over the past 40 years by the NSW Government. This
technology is not likely to be adversely affected if the climate changes as some
suggest. The work has implications for agricultural plant breeders and farmers both in
Australia and overseas, and there are substantial spillover benefits likely to accrue to
other industries and societies.
8
Table 1. Measured and calculated (climate data) yield increases due to the or
gene for NSW
Period
No. of trials
1980 - 1983
56
1997 - 2001
131
All
187
Long term
modelled
(a) With and without the or gene
Mean yield (t/ha) (a)
With
Without
3.36
3.04
3.56
3.37
3.46
3.21
Yield gain
t/ha
0.32
0.20
0.26
% increase
10
6
8
12
Table 2. Expected increase in production for NSW due to full application of the
or breeding technology
Average yield
Increase
Adjusted
2000-01 to
due to
increase(b)
2001-02 (a)
osmoregulation
t/ha
t/ha
t/ha
2.07
0.24
0.18
(a) Source: AWB website
(b) Adjusted for or gene frequency
Average area
of production (a)
2000-01 – 2001-02
‘000 ha
3696
Increased
production
‘000 t
665
Table 3. Summary of Australian wheat areas sown, production and yield by
GRDC Agroecological Zone, average of years 2000 – 2002, and percent yield
change due to or gene
Zone name
Area
Wheat
Average
Yield
harvested
production yield (c)
increase
(a)
(b)
due to or
gene (d)
Qld Central
NSW NE/QLD SE
NSW NW/QLD SW
NSW Central
NSW VIC Slopes
VIC HR/TAS
SA VIC BordertownWimmera
SA VIC Mallee
SA Midnorth-Lower
Yorke, Eyre
WA Sandplain
WA Northern Zone
WA Eastern Zone
WA Central Zone
51
842
844
1233
609
79
98
2001
1574
2980
1913
279
1.9
2.4
1.9
2.4
3.1
3.6
+30
+5
+20
+21
+12
0
486
1487
1458
2588
3.0
1.7
0
+50
787
261
1330
668
2029
2015
454
2031
954
3680
2.6
1.7
1.5
1.4
1.8
0
+66
0
+24
+3
22027
2.1
10704
(a) Units of ‘000 ha
(b) Units of ‘000 t
(c) Units of t/ha
(d) Percent change in yield
Source: ABARE surveys
Total
9
Table 4. Estimates of proportions of wheat production by pay grade for states
and Australia, based on 5 years to 2001
Wheat grade
Prime
Hard Premium Standard Other Total
hard
white
white
%
%
%
%
%
%
100
Queensland
34
43
15
7
2
Northern NSW
27
51
17
0
5
100
100
Southern NSW
6
36
25
27
6
100
NSW
14
41
22
17
6
100
Victoria
3
36
30
28
3
100
South Australia
3
37
32
26
3
100
Western Australia
0
37
35
27
1
100
Australia
8
39
28
22
3
Table 5. Summary of economic returns from wheat breeding with the or gene
NPV of benefits (2003 $’million)
Time to adopt
5 years
10 years
15 years
Ceiling
20%
50% 80% 20% 50% 80% 20% 50% 80%
NSW
Lag to
adoption
5
236
598
959
198
502
807 164 418
672
(years)
10
160
407
654
129
328
528 101 259
418
Australia
Lag to
adoption
5
898
2259 3621 754 1900 3045 627 1583 2538
(years)
10
611
1540 2470 492 1245 1997 388 984 1581
Table 6. Sensitivity analysis of economic results for Australia, lower production
figures and 8% yield advantage
NPV of benefits (2003 $’million)
Time to adopt
5 years
10 years
15 years
Ceiling
20%
50% 80% 20% 50% 80% 20% 50% 80%
Lag to
adoption
(years)
5
10
613
416
Lag to
adoption
(years)
5
10
483
327
1547
1054
1990s production base
2481 515 1301 2087
1692 335
851 1367
427
263
1083
672
1739
1082
1222
832
8% yield advantage
1961 405 1027 1649
1337 263
672 1082
336
206
855
530
1374
854
10
References
ABARE (1999), Australian Grains Industry: Performance by GRDC Agroecological
Zones, ABARE Report Prepared for the Grains Research and Development
Corporation, Canberra, April.
Alston, Julian M. (2002), ‘Spillovers’, Australian Journal of Agricultural and
Resource Economics, 46(3): 315-46.
Alston, Julian M., Marra, Michelle C., Pardey, Philip G., and Wyatt, T.J. (2000),
‘Research returns redux: a meta-analysis of the returns to agricultural R&D’,
Australian Journal of Agricultural and Resource Economics, 44(2): 185-215.
Alston, J.M., Norton, G.W. and Pardey, P.G. (1995), Science under Scarcity:
Principles and Practice for Agricultural Research Evaluation and Priority Setting,
Cornell University Press: Ithaca and London.
Commonwealth of Australia, Bureau of Meteorology (2003), Annual Climate
Summary 2002.
http://www.bom.gov.au/media releases/climate/change/20030106.shtml.
Cox, Thomas, Mullen, John and Hu, Wensheng (1997), ‘Nonparametric measures of
the impact of public research expenditures on Australian broadacre agriculture’,
Australian Journal of Agricultural and Resource Economics, 41(3): 333-60.
CSIRO (2001), Climate projections for Australia. CSIRO Atmospheric Research,
Melbourne. http://www.dar.csiro.au/publications/projections2001.pdf.
Gomez-Macpherson, H. and Richards, R.A. (1995), ‘Effect of sowing time on yield
and agronomic characteristics of wheat in south-eastern Australia’, Australian Journal
of Agricultural Research 46: 1381-1399.
Marshall, Graham R. and Brennan, John P. (2001), ‘Issues in benefit-cost analysis of
agricultural research projects’, Australian Journal of Agricultural and Resource
Economics, 45(2): 195-213.
Morgan J.M. (1977), ‘Differences in osmoregulation between wheat genotypes’,
Nature 270: 234-235.
Morgan J.M. (1980), ‘Possible role of abscisic acid in reducing seed set in waterstressed wheat plants’, Nature 285: 655-657.
Morgan J.M. (1983), ‘Osmoregulation as a selection criterion for drought tolerance in
wheat’, Australian Journal of Agricultural Research 34: 607-614.
Morgan J.M. (1991), ‘A gene controlling differences in osmoregulation in wheat’,
Journal of Plant Physiology 18: 249-257.
Morgan J.M. (1995), ‘Growth and yield of wheat lines with differing osmoregulative
capacity at high soil water deficit in seasons of varying evaporative demand’, Field
Crops Research 40: 143-152.
11
Morgan J.M. (1999), ‘Pollen grain expression of a gene controlling differences in
osmoregulation in wheat leaves. A simple breeding method’, Australian Journal of
Agricultural Research 50: 953-962.
Morgan J.M. (2000), ‘Increases in grain yield of wheat by breeding for an
osmoregulation gene: relationship to water supply and evaporative demand’,
Australian Journal of Agricultural Research 51: 971-978.
Morgan J.M. (2001), ‘Osmoregulation gene (or) frequency in Australian and other
wheat cultivars and accessions’, Proc. 10th Assembly of the Wheat Breeding Society
of Australia 10-13.
Morgan, J.M. (2002), ‘The drought tolerance gene in Australian wheat cultivars – an
overview’ in Martin, Bob (Ed), Research Update: Update of research in progress at
the Tamworth Centre for Crop Improvement 2001, NSW Agriculture.
Morgan, J.M. (2003), ‘Impact of climate change on crop water stress and yield
increases due to an osmoregulation (drought tolerance) gene in wheat’ in Martin, Bob
(Ed), Research Update: Update of research in progress at the Tamworth Agricultural
Institute 2002, NSW Agriculture.
Morgan J.M. (2004), ‘From process to production – the pursuit of understanding and
commercial utilisation of osmoregulation to increase yield of wheat and other crop
species under conditions of water stress’, Functional Plant Biology (submitted).
Morgan J.M., Hare, R.A. and Fletcher, R.J. (1986), ‘Genetic variation in
osmoregulation in bread and durum wheats and its relationship to grain yield in a
range of field environments’, Australian Journal of Agricultural Research 37: 449457.
Morgan, J.M. and Tan, M.K. (1996), ‘Chromosomal Location of a Wheat
Osmoregulation Gene Using RFLP Analysis’, Australian Journal of Plant Physiology
23: 803-806.
Powell, C. (2002), ‘Winter crop variety experiments 2001’, NSW Agriculture,
Orange.
12